Quick Take
Hertz’s rollout of AI vehicle scanners shows real promise for faster, more consistent inspections, but the tech is not fully baked. IT leaders should not blindly deploy this tech. Instead, they should require human-in-the-loop review and clear dispute workflows before any automated damage charge is finalized.
Why You Should Care
AI vehicle inspection can boost consistency and speed. Hertz partnered with UVeye to automate checks of body, glass, tires, and undercarriage, aiming for quicker turnarounds and better fleet availability. These benefits matter in any high-velocity operation. But “faster” only helps if decisions are trusted. Unfortunately, early customer reports flag reliability and governance gaps. Some renters describe charges for minor scuffs at premium rates and difficulty reaching a human to contest findings, eroding confidence and creating reputational risk. AI without appeal is just an automated argument. There is still a bright side. AI can identify safety issues (e.g., tire treads) in seconds, supporting preventive maintenance. The value case is there, but what’s missing is the governance wrapper: calibrated thresholds, audit trails, and customer-friendly recourse. Like most AI in the wild, this needs maturation, not abandonment.
What You Should Do Next
If you want to use this tech, you should just implement it where the AI proposes, but the humans decide. Do not deploy automated billing without human confirmation and photo evidence shared with the customer.
Get Started
- Define acceptance criteria, including lighting conditions, cleanliness requirements, and confidence thresholds per damage type. If any prerequisite fails, default to manual inspection. Tie criteria to measurable KPIs (e.g., charge reversals). If too many customers successfully dispute charges, that is a signal that the inspection criteria may be too aggressive or error-prone.
- Build a single source of truth evidence pack with time-stamped, angle-varied images and cost rationale auto-shared with customers. Unresolved cases escalate to a trained agent within 24 hours.
- Calibrate with A/B sites by running AI and human vs. human-only at matched locations. Compare turnaround, revenue leakage, dispute rates, and customer sentiment before scaling.